www.nature.com/scientificreports OPEN Subtractive inhibition assay for the detection of Campylobacter jejuni in chicken samples using surface Received: 5 March 2019 Accepted: 13 August 2019 plasmon resonance Published: xx xx xxxx Noor Azlina Masdor1,2, Zeynep Altintas3, Mohd. Yunus Shukor4 & Ibtisam E. Tothill1 In this work, a subtractive inhibition assay (SIA) based on surface plasmon resonance (SPR) for the rapid detection of Campylobacter jejuni was developed. For this, rabbit polyclonal antibody with specifcity to C. jejuni was frst mixed with C. jejuni cells and unbound antibody was subsequently separated using a sequential process of centrifugation and then detected using an immobilized goat anti-rabbit IgG polyclonal antibody on the SPR sensor chip. This SIA-SPR method showed excellent sensitivity for C. jejuni with a limit of detection (LOD) of 131 ± 4 CFU mL−1 and a 95% confdence interval from 122 to 140 CFU mL−1. The method has also high specifcity. The developed method showed low cross- reactivity to bacterial pathogens such as Salmonella enterica serovar Typhimurium (7.8%), Listeria monocytogenes (3.88%) and Escherichia coli (1.56%). The SIA-SPR method together with the culturing (plating) method was able to detect C. jejuni in the real chicken sample at less than 500 CFU mL−1, the minimum infectious dose for C. jejuni while a commercial ELISA kit was unable to detect the bacterium. Since the currently available detection tools rely on culturing methods, which take more than 48 hours to detect the bacterium, the developed method in this work has the potential to be a rapid and sensitive detection method for C. jejuni. It is estimated that the yearly medical and productivity losses caused by Campylobacter infections are over USD one billion1. Tere are more than 30 species and eleven subspecies in the genus Campylobacter. In poultry, the most ofen found species is C. jejuni, and it is the leading cause of foodborne illnesses in man. As chicken meat is the top consumed food globally, Campylobacteriosis risk is higher in chicken meat than in other meat and poul- try products2. Biosensor ofers a rapid, sensitive and robust detection methods for pathogen including C. jejuni3. Surface plasmon resonance (SPR) is widely used for the determination of a wide range of analytes, including bac- teria via direct detection. However, the direct detection of bacteria has some limitations; chiefy it is less sensitive due to the restricted efcient penetration degree of the evanescent feld coming up within the circumstances of total internal refection (TIR), which happens to be roughly 300 nm4,5. Bacteria, including C. jejuni with a size of around 5 µm, exceeds the evanescent feld limit. Tus, only a meagre measurable signal can be obtained from a small section of the bacterium5,6. To date, the most sensitive detection of C. jejuni with SPR platforms showed a limit of detection (LOD) value of 102 CFU mL−1 using the receptor binding protein (RBP) of the Campylobacter bacteriophage NCTC 126737 fol- lowed by using commercial polyclonal antibodies achieving a LOD of 103 CFU mL−1 6. A more recent SPR-based method for the detection of C. jejuni developed by our group yield a LOD value of 4 × 104 CFU mL−1 5. Although the frst method that relies on Campylobacter bacteriophage is sensitive, this bioreceptor is not commercially available, and its production requires a complicated procedure. Hence, the use of antibodies as bioreceptor for the development of C. jejuni detection is still a major choice for food samples analysis. However, due to the limitations arising from SPR penetration depth in case of detecting large pathogens, the results generally lack sensitivity. Te 1Cranfield University, Cranfield, Bedfordshire, MK43 0AL, England, United Kingdom. 2Biotechnology and Nanotechnology Research Center, Malaysian Agricultural Research and Development Institute, MARDI, P. O. Box 12301, 50774, Kuala Lumpur, Malaysia. 3Technical University of Berlin, Straße des 17. Juni 124, Berlin, 10623, Germany. 4Department of Biochemistry, Faculty of Biotechnology and Biomolecular Sciences, Universiti Putra Malaysia, 43400, Serdang, Selangor, Malaysia. Correspondence and requests for materials should be addressed to N.A.M. (email: [email protected]) SCIENTIFIC REPORTS | (2019) 9:13642 | https://doi.org/10.1038/s41598-019-49672-2 1 www.nature.com/scientificreports/ www.nature.com/scientificreports Figure 1. Schematic of the subtractive inhibition assay format. penetration depth generally does not allow performing sandwich assays with desirable LOD as it increases the height of the sensor surface further. An emerging technique to overcome this problem in SPR-based detection of C. jejuni is the subtractive inhi- bition assay. Tis method (Fig. 1) progress with an initial mixing of antibody and bacterial cells, followed by the separation of the unbound from the cell-bound antibodies via sequential centrifugation and fnally, the remaining unbound antibody is quantifed through the interaction with a previously immobilized anti-antibody on the SPR sensor chip surface8. As the size of the antibody is within the penetration depth of the evanescent feld, this boosts the level of sensitivity of the SPR for the detection of bacteria9–12. In the present work, a subtractive inhibition assay to develop a sensitive SPR-based immunosensor for the detection of C. jejuni using a rabbit polyclonal antibody with specifcity to C. jejuni is reported for the frst time. Highly sensitive and specifc quantifcation of this bacterium is successfully achieved using this approach to the best of our knowledge. Results Primary capture antibody concentration optimization. Te optimal concentrations of primary and secondary antibodies are vital to achieve a maximum binding response. In order to optimize this step, various concentrations of the primary antibody (goat F(ab) anti-rabbit IgG H&L antibody) at 50, 70, 100 and 150 µg mL−1 were frst immobilized on the SPR sensor chip followed by the injection of 100 µg mL−1 of the secondary a rab- bit polyclonal antibody with specifcity to C. jejuni. Te sensorgram in Fig. 2a shows that the immobilization response changed with the increasing concentration of the primary antibody, and Fig. 2b displays the binding response of 100 µg mL−1 secondary antibody at each primary antibody concentration used. From these results, the primary antibody concentration of 150 µg mL−1 gave the highest signal responses for both antibody immobiliza- tion and secondary antibody capture. ANOVA analysis of the results show that there was a signifcant diference (p < 0.05) in the immobilization response obtained from the primary antibody at 150 µg mL−1 compared to other antibody concentrations. Hence, 150 µg mL−1 was chosen as the optimum concentration for immobilization of the primary antibody. Higher concentrations of primary antibody were not tested as this would not be economical. In addition, the binding response obtained at 150 µg mL−1 was considered satisfactory. Further investigation of the injection period shows that a 4 min (100 µL) injection period gave a signifcantly higher response (p < 0.05) than a 3-min injection (75 µL). Hence, a 4 min (100 µL) injection was chosen for further optimization studies (Fig. 2c). Optimization of the free unbound secondary antibody separation technique. In a subtractive inhibition assay, either fltration or a centrifugation method with a single or a sequential step has been employed to separate the free unbound secondary antibodies. Tus, all of the techniques were explored. Te results in Fig. 3 show that the best technique giving the highest inhibition value (37.6 RU) was a sequential centrifugation step lasting for 2 min. Te second-best method was with the fltration step by employing a flter possessing a MWT cut-of of 0.2 µm, and this led to an SPR signal of 35.1 RU. Te lowest result was obtained with a fltration using a 0.1 µm MWT cut-of flter that provided an inhibition value of 4.2 RU. Optimization of secondary antibody binding to C. jejuni cells. The results in Fig. 4 show that 150 µg mL−1 was the best secondary antibody concentration exhibiting the highest binding response of about 100 RU at 5 × 107 CFU mL−1 concentration of C. jejuni. Te second-best concentration for secondary antibody was 125 µg mL−1 with an SPR signal of 59.5 RU at 5 × 107 CFU mL−1, which was about 40% lower than the response observed using 150 µg mL−1. Remarkably, this diference is even more pronounced at lower C. jejuni cell concen- trations. For example, at 5 × 103 CFU mL−1, the binding responses obtained with 150 µg mL−1 and 125 µg mL−1 secondary antibody were 70.74 and 11.8 RU, respectively. Tis is about an 83% reduction of the signal. SCIENTIFIC REPORTS | (2019) 9:13642 | https://doi.org/10.1038/s41598-019-49672-2 2 www.nature.com/scientificreports/ www.nature.com/scientificreports Figure 2. Sensorgram for optimization of antibody immobilization using diferent concentrations of primary capture antibody (goat F(ab) anti-rabbit IgG H&L antibody). (a) Sensor responses obtained due to the binding of 100 µg mL−1 of the secondary a rabbit polyclonal antibody with specifcity to C. jejuni. (b) Comparison of the binding response obtained from 3 and 4 minutes of injection period of the primary capture antibody during the immobilization process. (c) Error bars represent the mean ± standard deviation of triplicates. Direct subtractive inhibitive immunoassay for detection of C. jejuni. Te sensorgrams in Fig. 5a shows the direct binding of C. jejuni to the immobilized capture secondary antibody over the entire calibration range from 5 to 5 × 106 CFU mL−1 and control (PBS). Te data was transformed and normalized into the ratio 9 R/R0, which is each sample average response divided by the value of the average control .